Featured Tesla Autopilot recall probed by safety regulator following new crashes

Having been plagiarized, there are some posting styles that need to be shared:

• Identify cut-and-paste authors - you may agree with the quoted material but those are NOT YOUR WORDS. At a minimum enclose in quotation marks, parenthesis, or other delimiters and provide the source. My preference is indented and italics to make it clear I did not write those words.
• Your original content is important - I love to read your original words to understand where you are coming from. We can have an excellent dialog but only if I know whose words are which.
• Avoid quoting that changes the meaning - especially hard, I use ellipses, ". . .", to indicate where I have trimmed unnecessary words from a quote.

Many of these 'manifestos' suffer from word bloat by plagiarizing the words of others. Quotations can be useful if kept apart from your original content.

Bob Wilson

ps. I've been wanting to share this with our good Prius friends.

 

Evasion and Diversion as Manipulation Tactics (counsellingresource.com)
"Side-stepping and misleading: rather than being accountable and responsible, what the issue-dodger and subject-changer really wants is to advance their own agenda at the expense of yours"

 

Getting your agenda through is a multi-billion dollar industry.

 

Many people have been reading too much into some posts and becoming combative about the issue. So more detailed nuance helps there. The simple tldr is as follows

Tesla is popular and is an early provider of higher levels of autonomous and semi autonomous driving to more people

They've been increasingly involved in more accidents while these systems are active, with increasing numbers of serious injury and death

NHTSA has found that in many of these instances there was ample time for a person to intervene and either prevent or mitigate the collision/harm as a result of an unavoidable collision.

They are investigating if the systems to prevent/limit complacency and driver inattention are sufficient.

Tesla is the target of inquiry because the NHTSA is already investigating them for pushing an update last year to make the warning and inattention systems less intrusive and take longer to kick in. Another likely reason is that they *are* so popular and these advanced systems have been available to more people for longer through Tesla.

May point out that it is still a new technology - especially FSD - and that the naming conventions and the way the FSD and Autopilot systems are promoted and understood in the public consciousness likely doesn't help with the level of complacency build while using these systems.

NHTSA is, in effect, studying whether the driver attention systems in Tesla (and by extension others), need to be changed in order to reduce accidents that could otherwise be avoided. It's great if these systems reduce accidents overall. But if half of the accidents that *do* still happen have 5-10 seconds of lead time that a human could address but they don't because they're complacent/lazy, then reminders/checks/balances being used may not be sufficient.

Many have taken this to be a "Tesla bad everyone else good" targeted thing in the thread here. But really it's more like "Tesla the best case study" with snide commentary about how they can be flippant with a blowhard CEO about pushing things too fast for regulators.

 

Driver assist investigation is reminiscent of all the hubbub over car noise makers. We suddenly took focus on so many numbskulls not being able to be spatially aware of their surroundings. People walking into the pathway of the car. So suddenly the cure was implementing annoying background sounds ... so much so they became as ubiquitous as other background sounds.

More "Bell the Hybrid" nonsense | PriusChat

victim mentality
.

 

This is what I am curious about.
I have seen many reports stating the number of Tesla vehicles in accidents.
I have seen none comparing that to the number of non Tesla vehicles in accidents over the same time period (per mile driven).

 

You don't count "listening" as being part of "spatial awareness"? Also, visually impaired people exist and are allowed to go outside.

Would you disagree that ICE noise is ubiquitous, background, and also a useful tool in spatial awareness? Because all three of those things can be true at the same time.

 

What's too bad is you can't count the number of accidents it prevented because of it being active. Again, if you watch the weekly Wham Bam Tesla/Dash Cam YouTube channel, so many of those accidents could have been avoided if the the car was assisting in the driving instead of just having a zoned out driver at the wheel.

 

Sorry I guess I didn't clarify very well. Consider how some ice vehicles are just as quiet as an electric anyway. Why 'bell' just won mode of transportation.
It's irrational
Anyhow ...

- how confusingly surprising that Tesla would be in more accidents. I'm sure sales volume - outselling the ex #1 best selling car - the Corolla - has nothing to do with more accidents.

.

 

Two years ago, a Hyundai Elantra was idling so quietly in a parking lot that I had to stop and listen carefully to make sure it was really idling.

 

That's a good point. How you you prove that an accident was avoided when the reactions of the drivers of the other cars are not recorded, analyzed and correlated?

One of the youtube videos showed a Tesla using beta versions of the software driving in the wrong lane. When it entered a 4 lane road it made a left turn from the right lane in the middle of the intersection. How many cars had to take emergency actions to avoid the car that was misbehaving? How many navigation errors were not flagged as a problem during the "training" exercises? Chances are we will never know.

 

Tesla metrics: https://www.tesla.com/VehicleSafetyReport

Q4 2023

In the 4th quarter, we recorded one crash for every 5.39 million miles driven in which drivers were using Autopilot technology. For drivers who were not using Autopilot technology, we recorded one crash for every 1.00 million miles driven. By comparison, the most recent data available from NHTSA and FHWA (from 2022) shows that in the United States there was an automobile crash approximately every 670,000 miles.​

My 2019 Tesla Model 3 has 130,000 mi over 90% in either AutoPilot or Full Self Driving (FSD), I've witnessed how well they work in traffic. For example, FSD driving home from lunch:

• A to B :: car turned and moved to inside lane and the shorter stack of cars stopped at the light.
• B to C :: FSD stayed in lane and let another car change lane into the front.
• C to D :: in crowd of cars, FSD signaled, turned, and changed lanes to the outside lane and exit.
• D to E :: after turn into outside, 3d lane, the car changed to center lane before lane end.

Other than touching the volume control to signal I am paying attention, FSD performed perfectly.

Bob Wilson

 

Being better than an unaided driver isn't reason to not improve further.

 

The problem with Tesla's high level numbers is that it's not comparing like for like miles; that is, the autopilot numbers are likely to be highly skewed towards interstate driving, and of course the overall US numbers are for all types of vehicles and not just newer vehicles with ADAS. And they're talking autopilot and not FSD. In order to do a real statistical analysis, one would need much more detailed data on both sides of the equation. But that's why there exists such entities as the NHTSA, right?

 

Ugh, too many negatives, 'mental stack overflow:'

Did I parse it right:

• Being worse than aided driver is reason to improve.

If I got it right, the issue is evaluating "Being worse." My experience with the most recent FSD, it is equal to and sometimes better than this human driver . . . the one behind the wheel. Never forget that every driver believes they are better than all the rest.

Bob Wilson

 

A valid criticism, the FARS database provides a uniform set of collected data. I welcome anyone who has followed my practice of actually analyzing, not just speculating, the actual, fatal accidents. For example, I've often stated that "EVs are an applied IQ test" suggesting:

• EV owners are brighter and better drivers than the rest.

Bob Wilson

 

does fsd keep you at or below the speed limit when cars in front are driving recklessly?

 

Have you proven causation or even a correlation between IQ and driving ability?

(That's a joke. Do you have a link to you or anyone else making these analyses of more detailed data? Because I'm legitimately interested.)

 

Sometimes english sucks
I think the phrase meant was that FSD should continue to be improved even if it is equal or better than a driver without driver assist tech.

I don't believe anyone feels different. Continued improvement has always been the plan.

 

Zythryn got it.

I'll add that stating an ADAS is better than an unaided driver can be construed as deflection of a valid criticism of the ADAS.